Lexicometry: A Quantifying Heuristic for Social Scientists in Discourse Studies

Abstract

This chapter introduces lexicometry as a quantitative heuristic methodology for the analysis of discourses that complements qualitative hermeneutic methods. On this understanding, it draws a connection between Bachelard’s concept of ‘epistemic rupture’ and quantitative methods which allows the discovery of discursive phenomena prior to the interpretation of meaning in texts. Lexicometry is a corpus-driven approach that deploys, besides common corpus linguistic methods, complex algorithms to analyse the lexis of a given corpus exhaustively. It does so by contrasting different corpus parts organised in partitions. Taking examples from a corpus of 4000 press texts on the global financial crisis of 2008, the contribution illustrates how a large text corpus can be reduced systematically to a readable size. It also demonstrates different ways of exploring the lexicosemantic macro-structures using correspondence analysis, descending hierarchical classification, and other methods.

I am thankful to Malcolm MacDonald for his helpful comments on earlier versions of this text. Additionally, I want to thank André Salem for the numerous personal tutorial sessions and discussions of the software Lexico3 with which most of the analyses in this text have been conducted.